US 12,346,072 B2
Apparatuses systems and methods for optimization-based control of physical systems
Ankush Chakrabarty, Cambridge, MA (US); Rien Quirynen, Somerville, MA (US); Diego Romeres, Cambridge, MA (US); and Stefano Di Cairano, Newton, MA (US)
Assigned to Mitsubishi Electric Research Laboratories, Inc., Cambridge, MA (US)
Filed by Mitsubishi Electric Research Laboratories, Inc., Cambridge, MA (US)
Filed on Oct. 19, 2021, as Appl. No. 17/451,336.
Claims priority of provisional application 63/262,626, filed on Oct. 16, 2021.
Prior Publication US 2023/0119664 A1, Apr. 20, 2023
Int. Cl. G05B 13/04 (2006.01); G05B 13/02 (2006.01); G06F 30/27 (2020.01)
CPC G05B 13/048 (2013.01) [G05B 13/027 (2013.01); G05B 13/041 (2013.01); G06F 30/27 (2020.01)] 18 Claims
OG exemplary drawing
 
1. A predictive controller for controlling a system, the predictive controller comprising:
memory having instructions stored thereon; and
a processor configured to execute the instructions to cause the predictive controller, for each of control steps, to:
accept a feedback signal including a current output of the system indicative of a current state of the system;
classify the current output, using a trained function of the current output, for each solver in a set of solvers into one or multiple classes to produce a classification of the current output for each of the solvers, wherein each solver in the set of solvers is configured to solve an optimization problem by spending computational resources of the processor to produce a solution with an accuracy associated with the solver;
select a solver in the set of solvers based on one or a combination of the classification of the current output for the solver, the accuracy of the solver, and the computational requirement of the solver;
solve an optimal control optimization problem using the selected solver to produce a current control input, such that for at least some different control steps, the predictive controller solves a formulation of the optimal control optimization problem with different solvers having different accuracies, requiring different computational resources, or both; and
submit the current control input to the system thereby changing the current state of the system.